Jerry\'s IEEE EMCS Distinguished Lecturer Talk

1. IEEE EMCS Distinguished Lectures
“ Why Does My Product
Fail EMC ?”
And How To Fix it.
Jerry Meyerhoff
JDM LABS LLC
jerrymeyerhoff@ieee.org
http://rfjock.wordpress.com/
All Rights Reserved 2012

2. Outline / Overview
● Thanks For The Invitation !
● EMC Process Overview & Fundamentals <brief>
● Illustrate the 4 Quadrants Through Case Studies
● Problem Solving Methods: Technical & Humanistic
● Measurements, Experiments, Theory Formation, Modeling,
Solution Generation & Maintaining Progress
● Lessons-learned
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 2

3. EMC Tech Overview
EMC Definition
The Beneficial Property of a Device / System :
● Not to Create Excessive E-M Energy
● Not to be Disturbed by Expected E-M Energy
● Practical Design Goals & Regulatory Limits
● & Within Constraints
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 3

4. EMC Tech Overview
The 3 Players
● Source
● Path / Coupling Mechanism
● Affected Device or System / Receptor
-Choose the “Best” One(s) to Attack-
● Consider Constraints :
Performance, $, Time
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 4

5. EMC Tech Overview
SOURCE PATH RECEPTOR
Radio/TV/Cell wires Radio:AM/FM/TV
GPS Satellites cables Cell Hand-set
BlueTooth earth GPS -TomTom
High Voltage Lines metal BluetoothHeadset
Car building Car
iPOD person Pace-maker
Cell Hand-set air Hearing Aid
…
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 5

6. EMC Tech Overview
The EMC Matrix
Conducted Conducted
Immunity Emissions
Radiated Radiated
Immunity Emissions
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 6

7. EMC Tech Overview
4 Mechanisms + Plus One
Rank-order... Depends On Situation
● Conducted (mutual inductance)
● Magnetic Field
● Electric Field
● Electro-Magnetic Field (radio)
+ Path of Least Impedance
Re: Dr Tom VanDoren
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 7

8. EMC Case Studies
● Narrative Illustrations of ..
● Physical Principles
● Analysis Approaches
● Problem Solving
● Making Progress
● Working Well With Others
● “Best Solutions” Within Constraints
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 8

9. Process Words of Wisdom
Henry Ott's Big 4
1. Divide & Conquer : Structures, Freq's & Modes
2. Find The Dominant Mechanism : Pile On Fixes
3. Kill It Dead : Optimize Later
4. Implement Fixes : Best Possible Fashion
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 9

10. CE Case 1 Scanner 1
● The “Story”
● Large Table-top
Commercial
Document Scanner
● Goal : CE Mark
● Design Approach:
PC-AT Supply
CE-Marked
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 10

11. CE Case 1 Scanner 2
● Verification Test :
Fails CE @ Low MHz
● Fix : Add “Can” Filter
Pi+X/Y Caps
● So We're Done, Right ?
Uh-Oh, Not so Fast !
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 11

12. RE Case 2 Scanner Reprised 1
● Large Table-top
Commercial
Document Scanner
● The Plot Thickens
Way Beyond CE
● Boards & Lotsa Wires
● Sources, Paths &
Radiators
● Where 2 start ?
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 12

13. RE Case 2 Scanner Reprised 2
● Purist Find the Source
& Stop It There
● Close-field Probe
The “PC-104 Brains”
● Choke it Off 5X
● Not Enough !
● Keep Moving !
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 13

14. RE Case 2 Scanner Reprised 3
● I see 3 Antennas:
USB, FootSwitch, AC
● Measure Currents as
Lee Hill Advises
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 14

15. RE Case 2 Scanner Reprised 4
● Currents on 3
“Antennas”
● Same but different !
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 15

16. RE Case 2 Scanner Reprised 5
What to do ? All Three !
“Kill It Dead” H.Ott
● Move AC Filter. Isolate
& decouple wires
●
Got a “Free” 2nd Use for
Filter = TriZ principle
● Sounds easy now, but then?
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 16

17. RE Case 2 Scanner Reprised 6
● USB Gnd Floats in
Plastic
● USB GND bond demo
● Crude R/C filter in
Foot Switch
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 17

18. RE Case 2 Scanner Reprised 7
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 18

19. RE Case 2 Scanner Lessons 8
● Do not have to always contain it @ Source
PC104 cards: Just too many I/O & wires
● Thankfully found Dominant Effect (H-Ott)
Cables = Antennas & K-I-D @ low $
● Did not have to “predict” Pre-Compliance levels
Relative improvement @ Current Probe Enough
● Problems were Frequency & Source
independent / separable – Allowed independent
fixes applied w/o interactions , whew :)
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 19

20. CI Case 3 Load-Dump 1
● Vehicle Power Socket
Charger
● 24 V “Design”
● 600 Watt 39 V TVS
● Passed 12 Volt
“Load Dump”
● Verify 24 V Dump
110 ~ 170 Volts
200mS , 2~ 4 Ohm
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 20

21. CI Case 3 Load-Dump 2
● Monitor I & V
● Ramp Up Stress Volts
● 1.6 A slo-blow I2t=6.4
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 21

22. CI Case 3 Load-Dump 3
● 600 Watt TVS
● Clamps “soft”
● TVS Overheats to
SHORT
● Blows Fuse
● NOT Acceptable
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 22

23. CI Case 3 Load-Dump 4
● SPICE for Insights – kW , I^^2*t
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 23

24. CI Case 3 Load-Dump 5
● 5 kW TVS
● Clamps at Voltage
● Holds Voltage
● Clears Fuse, Slow I2t
● TVS Survives
● Change the Fuse OK
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 24

25. CI Case 3 Load-Dump Lessons 6
● Go / No-Go Testing Insufficient
● Gather “Real” Data While Verifying
● Understand the Physics .. Fuse I2t
● Simulate to Communicate
● Knowledge Makes it Easier to Sell Solutions
● Do Not Fear Going For The Max : 5kW
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 25

26. RI Case 4 - Big PCB 1
● So I Get an e-mail
● Commercial Vehicle
Control System
● Fails 3V/m @ 3 Labs
Copper Tape the Box
● Can I Fix It ?
● Client Visit - Overload
● Just Say Yes
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 26

27. RI Case 4 - Big PCB 2
Assessment
● ~15 cables (antennas)
● Plastic Box
● Huge PCB : Area & Thickness
● Multiple uP's & Analogue - Loosely Populated
● Minimal Failure Data – Something Resets
● “I'm the Expert” – Make Recommendations
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 27

28. RI Case 4 - Big PCB 3
How To Approach ?
● Study Schematic & Layout
● List of Best Design Practices & Why
● Pictures of PCB modifications
● Alternate PCB Stackups & Why: H Ott's Book
● Expectations Set : 3 Iterations
● Constraints : “If it took a little more $ to solve it
for The Holidays I'd be good with that ”
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 28

29. RI Case 4 - Big PCB 4
● PCB Stackup 4 layers
● 0.125” thick
● Signals Outers
● Power/Gnd Inners
● No Gnd Fills
● The Opportunity !
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 29

30. RI Case 4 - Big PCB 5
Marked-up PCB changes
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 30

31. RI Case 4 - Big PCB 6
● Found Allies : PCB Design Engineer & Ott Book
● Chapter 16 “The Rosetta Stone” of Stackups
● Bury Signals In Between Ground Planes
● Back-fill / Gnd-pour Open Signal Layer Areas
● Via-Stitch all Gnds Together Densely
● Bypass / Choke Every Cable (Antenna) Wire
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 31

32. RI Case 4 - Big PCB LESSONS 7
● Being There & Establishing Trust
● Just Say Yes .. Someone Else Would Too
● Find The Ally Within
● Use Common Language / Framework
● Ask For Everything & Early
● Communicate Frequently (Over Distance)
● Set Expectations & Limitations – Best Efforts
● Best Practices in Circuit & PCB, Work ...It Depends...
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 32

33. CE Case 5 – Hall Sensor 1
● Good IC RePackaged
● 2 terminals – simple !
● 500 Hz Pulses
Wheel Speed
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 33

34. CE Case 5 – Hall Sensor 2
● Just Measure CE in
AM Radio Band
● What The %#@&!
Over the Line ?
● OK 10nF Shunt
Cap Can Fix This
● & Need Cap for
ESD “2 for 1 Sale”
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 34

35. CE Case 5 – Hall Sensor 3
● Almost Better ...
● What's That Peak ?
● Find the Hidden
Resonant Circuit
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 35

36. CE Case 5 – Hall Sensor 4
SPICE Model For Insights
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 36

37. CE Case 5 – Hall Sensor 5
● LISN Gets Pulsed & Rung Out By SPICE
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 37

38. CE Case 5 – Hall Sensor 6
● FFT Peak ~500 kHz
● L-IS-N
● Impedance Stabilization?
● LISN Coils Resonance !
● 10nF & 10uH do the math
● ~500 kHz
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 38

39. CE Case 5 – Hall Sensor 7
● A Better Idea ? Use IC Application Circuit
● Do AC Analysis This Time
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 39

40. CE Case 5 – Hall Sensor 8
● Now 10 nF & 5 uH = 700 kHz By SPICE AC
● Do The Math, It Will Not Be Better
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 40

41. CE Case 5 – Hall Sensor LESSONS 9
● Beware of Fix's Unintended Interactions
● Industry Standard Fixtures Have Limitations
● Model All the Important Parts in SPICE
● As Simple As Possible But No Simpler (Einstein)
● Analyze By Inspection & Insight (resonance)
● Seek Corroboration : Analysis & Measures
“3's A Charm”
● Model When You Cannot Get In the Lab
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 41

42. RE Case 6 – Data RF Transmitter 1
● 433 MHz ISM Band
● Battery Op'
● 1 Chip Tx (PLL)
● uP : Tx Control, Data
● -20 dBc @1 MHz BW
● Official Measure
“Strange”
Non-compliant
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 42

43. RE Case 6 – Data RF Transmitter 2
● Measure Real-Time
● Late 1900's Tech ;-)
● Key-on Clicks & Pops
a'la HAM Radio
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 43

44. RE Case 6 – Data RF Transmitter 3
● Suspect Tx Enable
Time Delay < 1 mS
● Force Tx On Always
● Crude But Effective
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 44

45. RE Case 6 – Data Tx LESSONS 4
● Official Data Often Hides Causal Evidence
● Use Real-time Detection Equipment
● Read Component Spec's
● List Many Ideas & Dispatch Rapidly
● Do Quick 'n Dirty / Is-IsNot Experiments
● Software Was the Fix, For Once
● Experience Can Make It Look Easy
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 45

46. Summary of Summaries
● Just 3 Things: Source, Path, Affected
● 4 Problem classes RE, RI, CE, CI
● 4 Coupling modes - Tom V'
● 4 Tactics: Divide, Dominant,Dead,Impliment- Ott
● Understand Limitations
● Model Appropriately
● Best Practices Work – Up To a Point
● Human Factors as Important As Tech
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 46

47. Summary of Summaries , more
● Observe in Real-Time & Parametric Sweeps
● Look For Interactions & Independences
● Precompliance Relative Measures Work
● See The Antennas
● Make / Take Time to Organize & Communicate
● Just One More Thing ? (Colombo)
● Keep On Learning
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 47

48. Thanks For Giving Me
A Hand !
● Mentor: Yale Saffro, Bill Schilb, Gary Fierstein, Dick Bell
● Trainer: Ralph Mitchell, Michel Mardiguian, Henry Ott,Tom VanDoren,
Lee Hill
● Modeler: Nick Buris, Colin Brench, David Johns, Roy Leventhal
● Consultant: Nick Feagler, Daryl Gerke, Ken Wyatt, Patrick Andre'
● Business: Neil Hagglund, Jack Black, Tom Braxton
● Tech-Colleagues: Keith Hardin, Doug Smith
● EMCS: Bruce Archambeault, Kimball Williams
IEEE-EMCS-DL-1 jerrymeyerhoff@ieee.org 48